Profit Allocation and Corporate Taxing Rights: Global and Unilateral Perspectives Ana Seco Justo 1 Paris School of Economics - June, 2020 Supervisor: Thomas Piketty Referee: Gabriel Zucman Abstract By drawing from public data on the location of profits and economic activity of European Union multinationals (Foreign Affliate Statistics (FATS) and bilateral balance of payments), this paper provides an evaluation of the impact of the introduction of a formulary apportionment system and a global minimum effective corporate tax rate. Tax havens are the main losers from this transition to formulary apportionment. Developing countries can benefit from gains in tax base as large as €0.5 and €1 trillion, but this requires the introduction of population in the apportionment formula (rather than sales and/or employees). Otherwise, most benefits accrue to rich countries. This paper also looks at the possibility of unilateral strategies. If a country like France unilaterally adopts a minimum effective tax equal to 35%, including trade tariffs to multinationals operating at low-tax countries, I estimate that its tax revenues increase by a 2% of its national income (before possible retaliation strategies by other countries), as compared with a 1.4% under international cooperation. These estimates provide insights on the implications of the different proposed international tax reforms and shed light on tax policy design against profit shifting practices. In the absence of cooperation, it can be tempting for countries to adopt unilateral strategies to induce a move towards international cooperation. Keywords: Taxing rights, formulary apportionment, minimum tax rate, profit shifting JEL codes: H25, H26, F23, F68 1 I would like to thank both my supervisor Thomas Piketty and referee Gabriel Zucman for their guidance and useful advice. I would also like to give a especial thanks to my peers from PSE’s Master in Public Policy and Development (PPD) 2018-20.
Transcript
Profit Allocation and Corporate Taxing Rights:Global and Unilateral Perspectives
Ana Seco Justo 1
Paris School of Economics - June, 2020
Supervisor: Thomas Piketty
Referee: Gabriel Zucman
Abstract
By drawing from public data on the location of profits and economic activity of European
Union multinationals (Foreign Affiliate Statistics (FATS) and bilateral balance of payments),
this paper provides an evaluation of the impact of the introduction of a formulary apportionment
system and a global minimum effective corporate tax rate. Tax havens are the main losers from
this transition to formulary apportionment. Developing countries can benefit from gains in tax
base as large as €0.5 and €1 trillion, but this requires the introduction of population in the
apportionment formula (rather than sales and/or employees). Otherwise, most benefits accrue
to rich countries. This paper also looks at the possibility of unilateral strategies. If a country
like France unilaterally adopts a minimum effective tax equal to 35%, including trade tariffs to
multinationals operating at low-tax countries, I estimate that its tax revenues increase by a 2% of
its national income (before possible retaliation strategies by other countries), as compared with
a 1.4% under international cooperation. These estimates provide insights on the implications
of the different proposed international tax reforms and shed light on tax policy design against
profit shifting practices. In the absence of cooperation, it can be tempting for countries to adopt
unilateral strategies to induce a move towards international cooperation.
Profit shifting by multinationals creates sizeable tax revenue loses for non-haven countries. Recent
estimates indicate that, in 2015, 40% of multinationals’ profits were shifted to tax havens, generating
a global loss in tax revenues of 10% of the actual tax revenue collected (Tørsløv, Wier, and Zucman
(2018)). Following this research, non-haven European Union (EU) countries are the main losers from
this phenomenon, with a reduction in their tax revenue of around a 2%. Tax avoidance practices
also contribute to the rise of inequality between countries, by disproportionally undermining economic
resources for low-income countries, or more broadly developing countries, that are more reliant on this
source of revenues than advanced economies. The prospects of an increase in tax competition (IMF
(2019)) poses the question of the future of corporate income tax and the subsequent raise of profit
shifting practices.
As a response, corporate tax policy reforms have been proposed by both international institu-
tions and academics. The main claim is that the current international tax system, known as separate
accounting, is not adapted to the challenges that globalisation and digitalisation raises to multinational’s
taxation. A proposal, is the design of a taxation system that allocates taxing rights to countries where
multinationals’ economic activity is located2. This system is more commonly known as formulary ap-
portionment, and it has different variants which depend on the economic activity variables considered as
apportionment factor3, for example sales, or employees, or the definition of the tax base4. The objective
of it is to reduce the incentive of shifting profits to low-tax jurisdictions by allocating income where it
is generated. The OECD,5 with the Base Erosion and Profit Shifting (BEPS) project, the IMF, the
EU with the Common Consolidated Corporate Tax Base (CCCTB) or the Independent Commission for
the International Corporate Taxation (ICRICT) and academics as Avi-Yonah and Clausing (2007) have
advocated in favour of this taxation model. Another mechanism is the definition of a global minimum
corporate effective tax rate to limit international tax competition for attracting profits.2Countries as the United States, Canada or Germany have long relied in this type of systems to allocate corporate profits
within their territory. For example, the United States, as an answer to the high economic integration of the 50 states,
defined a system that apportioned profits based on multinationals’ location of property, payroll, and sales (Avi-Yonah and
Clausing (2007)).3
4For example, the IMF proposes an apportionment system based on the re-allocation of residual profit (i.e. the economic
rent derived by the multinational from its market power). This is known as residual profit allocation (Beer et al. (2020)).5The BEPS project is designed around two pillars. Pillar One refers to the need to reallocate profits to destination
markets. Pillar Two defines the implementation of a global minimum effective tax rate (OECD (2020)).
1
This paper aims to evaluate the associated outcomes from a transition to a formulary appor-
tionment system, or Global Formulary Apportionment (GFA) as defined in this paper, and the adoption
of a minimum effective tax rate, considering both the existence and absence of international coopera-
tion. For answering these questions, this paper draws from publicly available data on EU multinationals’
activity (Foreign Affiliates Statistics or FATS) combined with bilateral balance of payments statistics.
After a systematic correction of these statistics to allow comparability across them (see Section 3), the
resulting data set allows us to link the location of economic activity and pre-tax profits by multina-
tionals’ nation of origin. With this data set, the evaluation of the systems considered is focused on the
resulting profit allocation and the associated tax revenues under the different models. To assess the
true extent of profit redistribution, the analysis is based on a division of countries by both income and
development level. This approach is relatively novel in this literature. Cobham, Faccio, and FitzGerald
(2019) introduce it as well in their analysis of the redistribution of the tax base for US multinationals.
Related empirical evidence on formulary apportionment is mostly focused on the United States
thanks to publicly available multinationals’ data published by the Bureau of Economic Analysis (BEA)
and the Internal Revenue Service (IRS) (Cobham and Janskỳ (2019)). In the case of the EU, existent
evidence is limited and mostly for Germany (Fuest (2007)). Most of the studies rely on private data
sources as Orbis, or Amadeus at the EU level, from Bureau Van Dijk (De Mooij, Liu, and Prihardini
(2019)). The main drawback of these data sets is that companies’ subsidiaries data is incomplete6.
This paper contributes to this literature by providing evidence for the EU and by basing its results
on publicly available data. The advantages of applying these statistics, apart from their public status,
are their broader coverage as compared with private sources and their consistency with current profit
shifting estimates based on macroeconomic data (Tørsløv, Wier, and Zucman (2018)). However, these
data sets are not perfect and face different limitations. These limitations are in terms of their data
availability for low income and tax haven countries, or the existence of bilateral divergencies between
countries’ reported data (see Section 3 for a detailed discussion).
The definition of the apportionment factor, or economic activity variable, for a formulary
apportionment scheme and its both adequacy and redistributive impact, is one of the core questions
in this area of research. Cobham and Loretz (2014) asses the resulting redistribution across countries
by defining factors as tangible assets, the number of employees or their costs. The authors find that
if the number of employees is the unique apportionment factor, the highest redistribution is achieved.6Tørsløv, Wier, and Zucman (2018) provide a more in depth analysis of Orbis limitations. For example, in 2016 Apple
reported a global consolidated profit of 55.3 billion of EUR but the total sum by subsidiary just represented around a 4%
of the total.
2
Eichfelder, Hechtner, and Hundsdoerfer (2018) analyse the definition of payroll as a factor, following the
design of the German local business tax. An interesting insight of their research is the conclusion that
firms manipulate payroll expenses to save tax payments. The distortions introduced by this taxation
system have also been discussed by Hines Jr (2010) or Laffitte and Toubal (2019), in the case of tax
havens as sales platforms for United States multinationals. In this paper, three different factors are
considered: sales, number of employees and population. A tax system that defines multinationals’
profit allocation based on the share of population in each of the countries where it operates, has not
been considered before7. Empirical evidence points towards an inverse relationship between income
level and population (Alvaredo et al. (2017)). With population, a larger profit redistribution could
be achieved to countries at the bottom of the income distribution. This is particularly relevant for
policymakers, in the current context were the Coronavirus crisis is affecting disproportionally this set
of countries (Glassman (2020)).
The implementation of a global minimum effective corporate tax rate has been evaluated before.
Fuest, Parenti, and Toubal (2019) evaluate different tax policy proposals for France and conclude that
a worldwide minimum effective tax rate would be the best tool to contract profit shifting and raise
corporate tax revenues. In the same line, OECD research8 concluded that a minimum effective tax rate
could raise global tax revenues up to 4% of the current level and reduce incentives to shift profits due
to the elimination of tax differentials. A novel contribution of this paper to this area of research is the
economic assessment of a unilateral transition towards the adoption of a minimum effective tax rate for
a country like France.
The main results of this paper are that, if the EU were to adopt GFA system for its multi-
nationals, all income groups would be benefited from such a transition with increased taxing rights
and corporate tax revenues, especially in developing countries. I estimate that the tax base for these
countries increases between €0.5 and €1 trillion, but this requires the definition of population as an
apportionment factor (rather than sales and/or employees). This benefit is achieved at the expense of
tax haven countries9. The adoption of a minimum effective tax rate of 35% by a country like France,
derives sizeable tax revenues for the country, at a lower cost in terms of created import tariffs. As7In the United States, the federal income tax was apportioned across the different states according to their population.
In this case, tax liability was apportioned across states, i.e., contribution were made as a function of states population.
This mechanism was abolished in 1894 when it was declared unconstitutional (Saez and Zucman (2019)).8https://www.oecd.org/tax/beps/presentation-economic-analysis-impact-assessment-webcast-february-2020.pdf9See Table 18 for a complete list of the countries classified as tax havens in this paper. Note that, tax havens are not
differentiated in terms of income level. For example, Switzerland and Puerto Rico both fall in the same tax haven category
compared with an scenario with international cooperation, France derives higher tax revenue gains if it
decides to adopt a minimum effective tax rate unilaterally, 2% of its national income (before potential
retaliation strategies10 by other countries), as compared with a 1.4% under cooperation. Following that,
it can be tempting for countries to adopt unilateral strategies to induce a move forward to international
cooperation.
As it has been previously found in the literature, the final redistributive impact of formulary
apportionment is highly determined by the apportionment factor defined (Cobham and Loretz (2014)).
If the number of employees is considered, low income countries perceive a higher share of profits due
to multinationals outsourcing this activity to their territory. High income countries appear to be more
benefited under a system with sales due to their concentration of final consumers. Apportioning profits
by population achieves the highest redistributive outcome across income group. These results are
obtained in a framework that assumes no behavioural responses from both firms and governments to
a change in tax policy. As aforementioned, multinationals could manipulate the allocation of factors
to lower tax payments. In the case of sales and population, this may not be a concern due to the
immobility of the factors. However, it might not be the same for employees (Cobham, Faccio, and
FitzGerald (2019)) or even sales (Laffitte and Toubal (2019)). Empirical evidence based on the US
experience suggests that these behavioural responses could be limited (Clausing (2016)) or even result
in a null redistributive impact as compared with separate accounting (Altshuler and Grubert (2010)).
The rest of the paper is organized as follows. Section 2 formally describes the different tax mod-
els evaluated and their underlying mechanisms. Section 3 contains an explanation of the methodology
applied, the data corrections and sources for both EU multinationals’ corporate profits and economic
activity, and its descriptive statistics. Sections 4 and 5, present the results for both the GFA approach
and the adoption of a minimum effective tax rate, respectively. Section 6 concludes11.
2 International Corporate Taxation Models
This Section formally describes the international corporate taxation models evaluated in this paper: (i)
Separate Accounting (SA), (ii) Global Formulary Apportionment (GFA) and (iii) the mechanism for10As defined in Section 2, the proposed mechanism for the unilateral adoption of a minimum effective tax rate involves
the creation of trade tariffs. In this context, retaliation strategies refer to actions taken by countries so as to punish for
the costs levied on them. For example, reciprocal trade tariffs or other type of economic sanctions.11All the data and code needed to reproduce this paper can be found at https://www.dropbox.com/sh/fhyvy9xinumpe92/
the unilateral implementation of a minimum effective tax rate. Through the description of the different
alternative models, as in previous literature, the main assumption is the absence of behavioural response
from multinational enterprises (MNE) or governments to changes in tax policy.
Separate accounting (SA). The current international tax scheme, separate accounting,
treats each affiliate12 of a MNE as independent. Taxing rights are allocated to the jurisdiction of
residence of MNE’s affiliates (𝑎). Under this scheme, the corporate tax base for country (𝑖) of MNE
(𝑚), 𝑌 𝑚𝑖 , is equal to the sum of the positive pre-tax corporate profits 𝜋𝑎
𝑖 declared by affiliates (𝑎) of the
MNE at its territory,
𝑌 𝑚𝑖 = 𝜋𝑚
𝑖 =𝑁
∑𝑎
𝜋𝑎𝑖 (1)
where 𝜋𝑚𝑖 are the pre-tax profits registered by 𝑚 at 𝑖 with 𝜋𝑎
𝑖 > 0 and 𝑎 = {1, 2, 3…, 𝑁} is the
vector of affiliates. Defining 𝑡𝑖 (𝑡𝑖 ≥ 0) as country’s 𝑖 effective corporate tax rate, 𝑇 𝑚𝑖 = 𝑌 𝑚
𝑖 × 𝑡𝑖 is the
tax revenue collected by 𝑖 from the MNE (𝑚).
Global Formula Apportionment (GFA). Under GFA, the MNE is treated as a single
unit and its consolidated pre-tax profit Π𝑚 is considered. These profits are equal to the total income
generated by all the MNE’s affiliates, Π𝑚 = ∑𝐽𝑖 ∑𝑁
𝑎 𝜋𝑎𝑖 where 𝜋𝑎
𝑖 can be either positive or negative
and 𝑖 = {1, 2, 3, ...𝐽} denotes the vector of countries where affiliates are located. This implies that
the MNE can immediately offset losses in one affiliate against gains in others. By construction, this
so-called international loss consolidation leads to a reduction of the corporate tax base as compared
with separate accounting13.
In this framework, the tax base allocated to country 𝑖 is based on a weight or apportionment
factor (𝑤𝑚𝑖 ),
𝑌 𝑚𝑖 = 𝜋𝑚
𝑖 = 𝑤𝑚𝑖 × Π𝑚 (2)
12The term affiliate enterprise is broad concept that englobes subsidiaries, associates, branches or fellow enterprises
(OECD (2016b)). In this context, affiliates that operate at the country where the MNE is headquartered are also included.13For example, if a MNE registers 𝜋𝑚
𝐴 < 0 at country 𝐴 and 𝜋𝑚𝐵 > 0 at country 𝐵, under separate accounting the tax
base of the multinational is 𝜋𝑚𝐵 . However, if profits are consolidated, the tax base is reduced by 𝜋𝑚
𝐴 as profits of both
jurisdictions are aggregated. The impact of international loss consolidation has been previously discussed when evaluating
formulary apportionment methods. See for instance, Devereux and Loretz (2008) or Cobham and Loretz (2014).
5
with
𝑤𝑚𝑖 = 𝑧𝑚
𝑖𝑍𝑚
where 𝑧𝑚𝑖 = ∑𝑁
𝑎 𝑧𝑎𝑖 (𝑧𝑚
𝑖 > 0) and 𝑍𝑚 = ∑𝐽𝑖 ∑𝑁
𝑎 𝑧𝑎𝑖 (𝑍𝑚 > 0). In equation 2, 𝑌 𝑚
𝑖 corresponds
to the tax base allocated at country 𝑖 from 𝑚, which depends on Π𝑚, the consolidated profit and 𝑤𝑚𝑖 ,
the weight with 𝑤𝑚𝑖 ∈ [0, 1], 𝑧𝑚
𝑖 denoting the value of the factor registered by 𝑚 at 𝑖 and 𝑍𝑚 the total
global value of the factor for the MNE. As for separate accounting, 𝑇 𝑚𝑖 = 𝑌 𝑚
𝑖 × 𝑡𝑖 is the tax revenue
for 𝑖 arising from the MNE allocated profits at its jurisdiction.
In this framework, the parameter that determines the amount of profits allocated at 𝑖 is the
weight 𝑤𝑚𝑖 . It corresponds to 𝑖’s share of a factor related with the economic activity of the MNE at
its territory. In this paper, the GFA scheme considered defines weights based on both destination sales
and employees separately or an average of the two (ICRICT (2018)). With sales, the weight becomes
𝑤𝑚,𝑠𝑖 = 𝑠𝑚
𝑖 /𝑆𝑚 where 𝑠𝑚𝑖 are the sales of the MNE at country 𝑖 and 𝑆𝑚 = ∑𝑁
𝑖 𝑠𝑚𝑖 are global sales.
Equivalently for employees, 𝑤𝑚,𝑒𝑖 = 𝑒𝑚
𝑖 /𝐸𝑚 with 𝑒𝑚𝑖 as the number of MNE’s employees located at 𝑖 and
𝐸𝑚 = ∑𝑁𝑖 𝑒𝑚
𝑖 as the total labour force of 𝑚. A hybrid of both, is specifying 𝑤𝑚,𝑠𝑒𝑖 = 0.5×(𝑤𝑚,𝑠
𝑖 +𝑤𝑚,𝑒𝑖 )
and setting an equal weight for both factors. Population could also be defined as an apportionment
factor. This GFA scheme would correspond to a solidarity tax on MNEs profits (Chancel (2020))14. In
this case, the weight is defined as 𝑤𝑚,𝑝𝑖 = 𝑝𝑚
𝑖 /𝑃 𝑚 where 𝑝𝑚𝑖 is the population of country 𝑖 and 𝑃 𝑚 is
the total population where 𝑚 has registered profits
Note that the separate accounting system could be re-expressed under a GFA framework by
defining 𝑤𝑚𝑖 of Equation 2 equal to 𝑤𝑚,𝑆𝐴
𝑖 = ∑𝑁𝑎 𝜋𝑎
𝑖 /Π𝑚. A country 𝑖 will gain (lose) tax base with
a transition to GFA if 𝑤𝑚,𝑆𝐴𝑖 < 𝑤𝑚,𝐺𝐹𝐴
𝑖 (𝑤𝑚,𝑆𝐴𝑖 > 𝑤𝑚,𝐺𝐹𝐴
𝑖 ) where 𝑤𝑚,𝐺𝐹𝐴𝑖 is the weight considered
under GFA with the different apportionment factors (sales and/or number of employees and population).
Unilateral adoption of a minimum effective tax rate.15. To limit tax strategic setting
by countries, a global minimum effective tax rate has been proposed by international institutions as
the OECD (Pillar 2, OECD (2019)) or academics (e.g. Fuest, Parenti, and Toubal (2019)). As argued
by Saez and Zucman (2019), international cooperation might not be a requisite for its implementation.
Individual or a group of countries could engage in a unilateral transition to implement a minimum14At Chancel (2020), the idea of a European solidarity tax on MNE’s corporate profits is introduced. Its objective
is to finance Coronavirus related expenses for EU countries. In this work, the concept of European solidarity tax on
multinationals profits is adapted to a unitary taxation framework with population as the apportionment factor.15This model is based on the system described at Saez and Zucman (2019), Chapter 6. Note that, as mentioned by the
authors, this proposed model does not violate any existing international treaty (for example, double-taxation treaty).
6
tax rate. This mechanism would effectively tax MNE’s profits at a pre-defined minimum tax rate and
incentivize other countries to cooperate.
Define 𝑡∗ as the minimum effective corporate tax rate. If MNE’s affiliates (𝑎), register profits
at a jurisdiction (𝑖) and are taxed an effective tax rate 𝑡𝑖 lower than 𝑡∗ (𝑡∗ > 𝑡𝑖), a tax deficit is created.
This tax deficit (𝑑𝑚𝑖 ) is the corporate tax revenue not paid by affiliates due to its profits being allocated
at a tax jurisdiction with an effective tax rate lower than the minimum 𝑡∗.
Formally,
𝑑𝑚𝑖 =
⎧{⎨{⎩
(𝑡∗ − 𝑡𝑖) × ∑𝑁𝑎 𝜋𝑎
𝑖 , if 𝑡𝑖 < 𝑡∗
0, if 𝑡𝑖 ≥ 𝑡∗(3)
with 𝜋𝑎𝑖 > 0 and 𝑡𝑖, 𝑡∗ ≥ 0. From equation 3, the global tax deficit of the MNE is 𝑑𝑚 = ∑𝐽
𝑖 𝑑𝑚𝑖 .
This global tax deficit would be collected by an individual or group of countries (𝑐), that adopt the
role of tax collector of last resort and define 𝑡∗ as its effective tax rate. For MNEs headquartered at
its jurisdiction 𝑐 collects their entire tax deficit. For MNEs headquartered outside 𝑐, their global tax
deficit is allocated following a GFA formula. This system is described at equation 4.
𝐵𝑚𝑐 =
⎧{⎨{⎩
𝑑𝑚, if 𝑚 headquartered at c
𝑤𝑚𝑐 × 𝑑𝑚, otherwise
(4)
In equation 4, 𝐵𝑚𝑐 is the tax deficit collected by 𝑐 from 𝑚, 𝑑𝑚 is the global tax deficit of the
MNE and 𝑤𝑚𝑐 is the weight defined as in equation 2. This weight could be a function of sales (𝑤𝑚,𝑠
𝑐 ),
employees 𝑤𝑚,𝑒𝑐 or both (𝑤𝑚,𝑠𝑒
𝑐 ). A simpler version of this system is to allocate the global tax deficit
independently of where the MNE’s headquarters are located. Note that under this scheme, the separate
accounting system is maintained, the only modification is that an individual or group of countries collect
the MNE’s global tax deficit. Therefore, tax revenues under this scheme are 𝑇 𝑚𝑐,𝑢 = 𝐵𝑚
𝑐 + 𝑡∗ × 𝜋𝑚𝑐 , as
profits booked at country 𝑐 are taxed at the minimum effective tax rate 𝑡∗ and are not part of the global
tax deficit of the MNE. This assumes that if the country where to adopt the minimum effective tax
rate, profits allocation will not be altered as compared with the current international tax system. Note
that, as 𝑡∗ > 𝑡𝑖 by definition, a country will always have the incentive to adopt this mechanism in front
of a separate accounting system.
7
Following this system, the tax deficit could also be collected by imposing an sales tariff (𝜏𝑚𝑠 )
(or more broadly imports) and a per employee tax (𝜏𝑚𝑒 ) to the so-called tax deficitary MNEs. The tariff
on imports is equal to 𝜏𝑚𝑠 = 𝑇 𝑚
𝑐 /𝑠𝑚𝑐 , where 𝑇 𝑚
𝑐 corresponds to MNE tax deficit collected by 𝑐 and 𝑠𝑚𝑐 is
the value of imports from 𝑚 registered at 𝑐. In other words, the MNE will have to pay a tariff rate equal
to 𝜏𝑚𝑠 for its sales at 𝑐 due to its activity at low-tax jurisdictions. Re-arranging terms16, 𝜏𝑚
𝑠 = 𝑑𝑚/𝑆𝑚.
Therefore, 𝜏𝑚𝑠 is increasing with the global tax deficit of the MNE (𝑑𝑚), as the tax collected will be
higher, and it decreases the larger its total value of sales (𝑆𝑚), as the fraction apportioned to 𝑐 would
be lower. Equivalently, in the case of employees, 𝜏𝑚𝑒 = 𝑇 𝑚
𝑐 /𝑒𝑚𝑐 , where 𝑒𝑚
𝑐 are the total number of
employees registered by the multinational at 𝑐.
This unilateral adoption contraposes with the implementation of a minimum effective tax rate
under international cooperation. In a scenario without tax differentials where all countries cooperate
and implement a global minimum tax rate, profit shifting incentives will disappear. Following Guvenen
et al. (2017), the tax base under a GFA scheme is defined as the profit allocation in absence of profit
shifting17. In this cooperative scenario, and following previous notation, the tax revenue for country 𝑖 is
𝑇 𝑚𝑖,𝑐 = 𝑡∗ × Π𝑚 × 𝑤𝑚
𝑖 . An important question arises: When will a country 𝑖 have the incentive to adopt
a unilateral a minimum effective tax rate rather than cooperate for its international implementation?
We assume that country 𝑖 will decide its strategy based on the associated tax revenues gains for each
scenario with respect to a separate accounting system. Therefore, for country 𝑖 to adopt the unilateral
strategy the following condition needs to hold18
𝐵𝑚𝑖 ≥ 𝑡∗(Π𝑚 × 𝑤𝑚
𝑖 − 𝜋𝑚𝑖 ) (5)
From Equation 5, as long as the collected tax deficit by country 𝑖 from multinational 𝑚, 𝐵𝑚𝑖 ,
is larger than tax revenue gains from international cooperation, country 𝑖 will have the incentive to
unilaterally adopt a minimum effective tax rate rather than cooperate.16This expression is obtained by replacing 𝑤𝑚,𝑠
𝑐 by 𝑤𝑚,𝑠𝑐 = 𝑠𝑚
𝑐 /𝑆𝑚. Note that its corresponds to a system with just
sales as a weighting factor. If sales and employees are the weights, 𝜏𝑚𝑠 = 0.5 × (𝑑𝑚/𝑆𝑚) and 𝜏𝑚
𝑒 = 0.5 × (𝑑𝑚/𝐸𝑚) as
𝑤𝑚,𝑠𝑒𝑐 = 0.5 × (𝑤𝑚,𝑠
𝑐 + 𝑤𝑚,𝑒𝑐 ).
17The idea is that, if there are no tax differentials that may incentivize profit shifting, profit location should be aligned
with economic activity. Note that, as argued by Guvenen et al. (2017), this procedure assumes return differentials of
productive factors across locations. If not, tax base allocation would be equal for all countries.18Equation 5 is derived from: (𝑡∗𝜋𝑚
𝑖 + 𝐵𝑚𝑖 ) − 𝑡𝑖𝜋𝑚
𝑖 ≥ 𝑡∗(Π𝑚 × 𝑤𝑚𝑖 ) − 𝑡𝑖𝜋𝑚
𝑖 where 𝑡𝑚𝑖 𝜋𝑚
𝑖 is the tax base for 𝑖 under a
separate accounting system (see Equation 1).
8
3 Methodology and data
In this Section, the methodology applied to build up the data set for the economic activity and pre-
tax profits of EU multinationals is explained. Then, the different sources, their limitations, and the
procedures to overcome them are described.
3.1 Methodology
The objective of this paper is to evaluate the different models described at Section 2 for EU multi-
nationals. The evaluation is focused on the static distributional impact of a transition from separate
accounting to GFA, in terms of tax base re-allocation and tax revenues. For the adoption of a minimum
tax rate, the analysis is on EU’s MNEs global tax deficit and the associated tax revenue gains.
The objective is to measure pre-tax corporate profits of EU’s MNEs at the different locations
of their affiliates. For that, pre-tax corporate profits of an MNE (𝑚) at country 𝑖 are decomposed in
three different components (i) net dividends paid (𝐷)19, (ii) reinvested earnings (𝑉 ) and (iii) corporate
income tax paid (𝑇 ): 𝜋𝑚𝑖 = 𝐷𝑚
𝑖 +𝑉 𝑚𝑖 +𝑇 𝑚
𝑖 where 𝑋𝑚𝑖 = ∑𝑁
𝑎 𝑋𝑎𝑖 with 𝑋 = {𝐷, 𝑉 , 𝑇 }. This is what the
Corporate Income Tax (CIT) aims to tax. Dividends and reinvested earnings are the equity income that
correspond to the the MNE for its investment on the affiliate. For the purpose of this paper, corporate
income taxes paid by affiliates are inferred with available data for countries’ effective tax rates (ETR).
The final dataset for 2015, contains the pre-tax corporate profits of EU’s MNEs affiliates, the
value of their sales per destination, the number of employees hired at each affiliate and an estimate of the
corporate taxes paid. The data set is tabulated by country of location of the multinational headquarters
or the so-called parent and its different affiliates.
3.2 Data on Affiliates Economic Activity
Foreign Affiliates Statistics (FATS). FATS describe the economic activity of non-financial corpo-
rations’ affiliates. It reports indicators as value added, gross operating surplus, personnel costs, sales,
or the number of persons employed by the affiliate. Statistics are classified by the country of residence
of the affiliates’ control unit. This control relationship is defined when the control unit holds a majority19Net dividends paid are the difference between the dividends paid out by the affiliate to its parent and the dividends
received by the affiliate from its parent.
9
(more than 50%) of the voting power or the shares. Data is reported in a directional basis, either
inward or outward FATS. Inward FATS correspond to the economic activity of foreign affiliates within
the reporting economy. Outward FATS describe the economic activity of affiliates controlled by the
reporting economy outside its frontier. For example, defining Spain as the reporting economy, sales
of French affiliates in Spain would be recorded as inward FATS while the sales of Spanish affiliates in
France would be recorded as outward FATS.
The main data source for EU countries are the FATS published by Eurostat (Eurostat (2012)).
For non-EU countries, data is from the Activity of Multinational Enterprises (AMNE) database, which
is the OECD’s equivalent of FATS. For EU countries, the data from Eurostat’s FATS and AMNE
statistics are equivalent. Sales and employees data on outward FATS for EU corporations foreign
affiliates is used to build up the different weights (𝑤𝑚,𝑠𝑖 , 𝑤𝑚,𝑒
𝑖 and 𝑤𝑚,𝑠𝑒𝑖 ) for the GFA approach (see
Section 2). For comparability across the different international taxation models, just data for affiliates
with both sales and employees data available is considered. This implies a reduction in the amount of
profits of approximately a 4%. FATS do not report data of the activity of the MNE where its control
unit is located, for example, data on the economic activity of Spanish multinationals in Spain. It is
inferred by applying the share of sales and employees of local US MNEs with Internal Revenue Service
(IRS) data to the local sales and employees of the total EU corporate sector.20
These statistics suffer from two main limitations. First, the existence of bilateral discrepancies
between the data reported by the reporting and its counterpart economy. For example, in the case of
sales, EU and OECD partners recorded sales from UK controlled affiliates €422 billion larger than what
the UK reported21. Second, the incidence of unrecorded or unallocated values. In the case of Sweden,
aggregated sales by partner just represents a 31% of total sales, or in the case of Luxembourg, a 13%.
In order to reduce the impact of this limitations, EU countries outward data is corrected with mirrored
EU and OECD partners inward data22.20Data is from the Internal Revenue Service (IRS) https://www.irs.gov/pub/irs-soi/16it01acbc.xlsx. Data for the total
local employees and sales for the United States is from OECD’s inward FATS. The share of US local MNE sales is equal
to 37% of total local sales, and the equivalent for employees is 14%. Data for the EU corporate sector is from Eurostat’s
inward FATS.21See Table 26 at the Annex.22See Table 27 and 28 at the Annex for more details.
Foreign Direct Investment (FDI) statistics. FDI statistics contain both investment position and
its associated income and financial flows. Investment position is divided into equity and debt. Dividends
and reinvested earnings correspond to the generated flows from an equity position, and interests are
the equivalent for debt. Statistics are classified by geographical allocation of the immediate investor.
They are recorded when a cross-border investment of a resident investor (or parent) in a non-resident
business enterprise (or affiliate) occurs, and vice versa. According to the direction of the investment, data
is classified as inward or outward FDI. For example, inward FDI would correspond to the investment
reported by an affiliate located in Spain received from a French parent whereas outward FDI would
be the opposite, investment reported by a French parent in a Spanish affiliate. For the parent-affiliate
relationship to be established, the parent needs to have more than 10% of voting power of the affiliate.
FDI data is from three different sources: OECD, Eurostat, and IMF. For European Union
countries that are members of the OECD, data is from the OECD’s database on FDI statistics (OECD
(2009)). In this case, priority is given to OECD data in front of Eurostat due to the unavailability of
data for Special Purpose Entities (SPEs) 23 at the later source. Research indicates that both OECD
and Eurostat are generally very consistent (OECD (2016a)). For non-OECD European Union mem-
bers (Bulgaria, Cyprus, Malta, Romania and Croatia) data is from Eurostat tables bop_fdi6_pos and
bop_fdi6_inc. When equity position data is not available at either OECD or Eurostat, data is from
IMF’s Coordinated Direct Investment Survey (CDIS). Equity income on outward FDI at a by partner
level is used to derive the after-tax corporate profits of European MNE’s affiliates by aggregating both
dividends (𝐷𝑚𝑖 )24 and reinvested earnings (𝑉 𝑚
𝑖 ).
For the purpose of this paper, relying on FDI statistics faces different limitations. First, the
incidence of confidential or unallocated data on by partner FDI statistics. This is particularly relevant
for EU tax havens and low income countries.25 For example, in the case of Cyprus, aggregated by23An SPE is (i) A formally registered or incorporated legal entity that is resident in an economy and recognised as an
institutional unit with little or no employment (up to a maximum of 5 employees), little or no physical presence, and
little or no physical production activities in the host economy, (ii) Directly or indirectly controlled by non-residents, (iii)
established to obtain specific advantages provided by the host jurisdiction and (iv) transacting almost entirely with non-
residents with large parts of the financial balance of a cross-border nature (IMF Task Force of the Balance of Payments
Committee (BOPCOM))24Note that these dividends are gross of any withholding taxes.25See Table 23 at the Annex for more details. In the case of Luxembourg and Spain, countries do not publish by partner
data, just the aggregate value.
11
partner equity income flows just represent a 6% of the total world published. Second, there are bilateral
discrepancies between the data reported by an investor in an affiliate country and what the affiliate
has reported to receive, and vice versa. For example, Netherlands reported to receive €1.15 and €2
billion higher equity income from Switzerland and the United States, respectively, that what these
countries have recorded to pay26. Third, FDI statistics could over-estimate corporate profits due to its
classification by geographical allocation. As mentioned by Wright and Zucman (2018), to the extent
that multinationals use intermediate holding companies located in tax havens, FDI statistics are over-
estimated in such locations. For example, corporate profits of Amazon could be included on the outward
FDI income of Ireland, under a classification by geographical location, but in reality, they accrue to
United States, that is where the multinational is headquartered. To overcome the first two limitations,
equity income statistics are corrected in a bilateral basis with mirrored inward data27 for EU and
OECD partner countries28 and missing values are imputed with estimated returns on investment and
FDI position data29. To reduce the incidence of inflated FDI income due to its geographical allocation,
whenever data is available, FDI statistics are systematically corrected by subtracting SPEs values. SPEs
are commonly used to channel investment before reaching its final destinations (OECD (2016a)), and
therefore, eliminating them would ameliorate the distortion.
Note that FATS and FDI statistics are complementary, however differences in methodology
hamper a liable comparison (OECD (2005)). FATS statistics only cover a subset of the corporations
included in FDI as they just describe the activity of non-financial corporations. In addition, FDI
statistics are classified by the geographical allocation of the immediate parent, while FATS are classified
by the residence of the control unit. Following Tørsløv, Wier, and Zucman (2018), the only systematic
correction done to overcome the comparability barriers is to clean FDI data from SPEs, which are not
covered by FATS. As the authors argue, the margin of error should be relatively small as there is a
sizable overlap between multinationals and FDI firms.26See Tables 22 and 21 at the Annex for more details.27Following the previous example with France and Spain, from an outward FDI perspective, the mirrored data for France
is the inward data received from France as reported by Spain. Note that, in an ideal setting, both values should coincide.28See Table 25 for more details on the resulting corrected data.29In this context, return to investment (for either inward and outward FDI) corresponds to the ratio between the FDI
income flows and the corresponding equity or debt position (see Table 24 at the Annex). Due to this imputation, total
equity income increases in a 3%. Note that values are just imputed for observations that contain a positive value for sales
but no equity income data reported.
12
3.4 Additional data sources
Structural Business Statistics (SBS). SBS from Eurostat contains data of the economic activity of
companies with an European residence. It includes indicators as gross operating surplus, value added,
number of employees or turnover. As compared with FATS, SBS contains data for all the resident
companies within a territory while FATS is classified by residence of the control unit. For example, the
SBS of Spain will contain the economic activity of all firms located in Spain. On the other hand, FATS
will report statistics for just Spanish owned companies. SBS data is used to analyse the impact of the
ownership structure on both sales and employees statistics30.
Corporate tax rates. Statutory corporate tax rates are from KPMG and whenever they are
not available, data is from per country reports of World Bank’s Doing Business31. Effective Tax Rates
(ETR) for non-European countries, data is from Tørsløv, Wier, and Zucman (2018) or imputed with
Other data as corporate tax revenues is from OECD and IMF WoRLD, net national income
and population statistics are from the World Inequality Database (WID) and income and development
classification are from World Bank and United Nations Statistics Division (UNSTAT), respectively.
3.5 Descriptive Statistics
Table 1 presents the summary statistics of EU MNEs for 2015. EU MNEs registered a total consolidated
profit of €1.5 trillion with nearly 94% at high income (70%) and tax haven (24%) countries. This
represents a 15% of the corporate profits generated across the globe33. Corporate income taxes were
a 15% of pre-tax corporate profits, accrued principally to high income countries. At tax havens, this
ratio was of 10%, indicating that profits located there were taxed by a lower average effective tax rate.
In global terms, taxes paid represented a 12% of world corporate income taxes34.30See the Annex Section 8.2 for a more complete discussion.31https://www.doingbusiness.org/content/dam/doingBusiness/media/Annual-Reports/English/DB15-Full-Report.pdf32Imputation is based on the average ratio between effective and statutory corporate tax rate per development level
(i.e. developed, developing and transition) and tax havens. Average ratio per group: (1) developed = 74%, (2) developing
= 60%, (3) transition = 74% and (4) tax havens = 28%. The imputation process consists in the product between the
average ratio and the statutory corporate tax rate.33The total value of global pre-tax corporate profits for 2015 is approximately of €10 trillion. Source: Tørsløv, Wier,
and Zucman (2018) (Table 1)34Total corporate income taxes in 2015 are estimated to be equal to €1.9 trillion. Source: Tørsløv, Wier, and Zucman
(2018) (Table 1)
13
Economic activity was mostly concentrated at high income countries that accounted for a 80%
and 70% of total sales and employees, respectively. For countries at the middle and the bottom of the
income distribution, the ratio of employees with respect to the total is larger than for sales. For example,
1% of affiliates sales are done at lower middle-income countries while a 5% of their employees are hired
there. This indicates that at these countries the prevalent activity is employment. In respect to the
location of economic activity and profits at mid and low income, the total fraction of sales and employees
is generally larger than the one for profits. In the case of upper middle- income countries, while 17% of
total MNEs employees are at their territory around 5% total profits were booked there. The opposite
phenomena is observed at tax havens, where the total weight in economic activity, 11% for both sales
and employees, is lower than its share in total profits, of approximately a 24%35. Lower-middle- and
low-income countries concentrate half of the total population but nearly 0.12% of profits, while for high
income and tax havens population accounts for a 16% and recorded profits a 95% of the total36.
4 Redistributive Implications of Global Formulary Apportionment
In this section, the resulting profit redistribution under a GFA system as compared with separate
accounting is analysed. This analysis is done with both an income and development level perspective.
Table 2 and Figures 1 and 2 present the profit allocation under the different GFA systems considered
(sales, employees, sales and employees and population) classified by income level. From Table 2, for all
the models, tax havens are the main losers of a transition from a separate accounting system to GFA.
On average, tax havens experience a reduction of more that half their initial tax base. Conversely, the
mid and the bottom of the income distribution countries, concentrate the gains from such transition.
Taxing rights at high income countries remain at a similar level under the different GFA schemes as
with the current international tax scheme. The same pattern arises when country classification is done
by development level. From Table 3 and Figures 1 and 2, transition and developing countries gain tax
base under all GFA schemes while developed are positively affected at a smaller scale, expect when
considering population as a weight where this set of countries lose tax base. The gains for developing
countries are in between €0.5 and €1 trillion, depending on the GFA design considered.35This fact is documented as well by Garcia-Bernardo, Jansky, and others (2019) for US multinationals, were they show
that lower effective corporate tax rates are associated with higher levels of profits reported if compared with measures of
real economic activity.36Research has previously documented the existent inverse relationship between income and the level of population. See
for example, Figure 5 at Alvaredo et al. (2017) and Table 17
14
The distribution of the EU MNEs consolidated profits of €1.5 trillion depends on the weight
defined for each of the schemes considered. From Table 2, comparing between a GFA scheme appor-
tioning by sales or the number of employees, the later is a factor that favours the allocation of the tax
base towards countries at the mid and the bottom of the income distribution. With a GFA scheme
weighted by sales, high income countries increase their tax base in a 14% and tax havens experience the
lowest contraction of their tax base, of 60%. Lower middle- and low-income nearly double it, and with
employees it is more than tripled reaching around €0.7 trillion. In general terms, this change implies
that the share of EU MNEs profits allocated at lower middle and low-income countries increases from
a 0.93% of the total to a 1% and 5% when weighting by sales or employees, respectively (see Table
1). When considering both sales and employees, the resulting distributional impact is in between the
results obtained when both weights are considered separately. This is due to the weight design under
this scheme. With population as a weight, the 95% of profits allocated to high income economies and
tax havens under separate accounting (see Table 1) drops to 16%, leaving the rest allocated to mid and
low income economies. Tax havens register the higher lost, with their tax base practically re-allocated
out.
To understand the economic implications of this tax base redistribution, it has to be translated
to tax revenues. As in related literature, the assumption is that re-allocated profits would be taxed
at the statutory corporate income tax (CIT) (Cobham, Faccio, and FitzGerald (2019)). In spite of
international loss consolidation that reduce the tax base in approximately a 2% (see Table 2), from
Table 4, tax revenue under all the GFA schemes is larger than under separate accounting.
Figure 3 shows the resulting tax revenues arising from the activity of EU MNEs per income
and development level expressed as a percentage of net national income. Tax haven countries derive
a substantial benefit from the current international tax scheme. This set of countries collected tax
revenues from EU MNEs representing approximately 4% of their national income, three percentage
points larger than in the case of high-income countries. This striking result is in line with the finding
by Tørsløv, Wier, and Zucman (2018), that relate this phenomenon to the ability of havens to attract
foreign firms’ profits. Even if, from Table 2, tax havens lose more than half of their tax base under the
different GFA schemes, this group is still the one with the highest ratio of revenues-to-national income,
of approximately a 2% in the case of defining a weight with sales. As aforementioned, under a GFA
scheme with sales, tax havens experience the lowest reduction in tax base, of nearly a 60%. What are
the reasons behind this phenomenon? Research documents that the establishment of sales platforms in
tax havens serves multinationals to shifts profits (Laffitte and Toubal (2019)), therefore, sales at their
15
territory could be more aligned with the amount of booked profits37. Another potential explanation
is that big EU economies are included under the category of tax havens as, for example, Netherlands
or Belgium. Therefore, an important share of final consumers is concentrated at their territory, and
thus sales. However, this should also be reflected with larger revenue gains for high income economies,
which, as it can be seen in Figure 3, is not the case.
As from Table 4, high income countries tax revenues as a percentage of national income remain
stable across the different systems, at around a 1%. The mid and the bottom of the income distribution
experience gains in terms of tax revenue, with an average gain between 0.2 and 0.3 percentage point for
GFA schemes with sales and/or employees. Therefore, although tax base gains are large, tax revenue
gains are comparatively modest. Under a GFA with population, tax revenues to national income are
the highest for this set of countries arriving to a 6% in the case of low income, nearly a 6 percentage
points increase. Under separate accounting, tax revenues from EU MNEs yield around €1,348 per
capita for tax haven, €260 for high-income, and for the rest an average of €3 per capita. This ranking
is maintained through the different GFA schemes. If a weight with population is defined, transfers in
per capita terms is are harmonized across groups, with an average transfer of €60 per capita. Table 5
presents the same conclusion but with a country classification based on development level. In this case,
on average, tax revenues for developing countries raise by a 0.5% of their national income.
However, note that the above presented results are derived from a framework that does not
integrate the behavioural response from firms and governments from a change in tax policy. A GFA
scheme that includes employees may discourage the allocation of this activity at high-tax jurisdictions
in favour of tax havens, and the same phenomena could happen in the case of sales (Laffitte and Toubal
(2019)). There exists a relatively tiny body of research that aims to integrate such responses through a
general equilibrium model (Fuest, Parenti, and Toubal (2019)) which is not considered here. Previous
research in this are suggests that the incidence of behavioural responses in this results may be limited
(Clausing (2016)) or neutralize the total redistributive impact as compared with separate accounting
(Altshuler and Grubert (2010)).37As explained at Section 3 sales data is from FATS. In this data set, sales are defined as market sales of goods and
services to third parties. Therefore, the possibility of recorded sales with profit shifting objectives is not ruled out.
16
5 Unilateral Scenario: Adopting a Corporate Minimum Tax Rate
As described at Section 2, the global tax deficit is equal to the missing tax revenue due to profits
being located at jurisdictions with a tax rate below a pre-defined minimum effective tax rate. Following
that, Table 6 shows the global tax deficit for different levels of effective tax rates, or ETR, for MNEs
headquartered at the EU and at the rest of the world. Note that, for non-EU MNEs, the tax deficit
just covers lost revenues from their booked profits at the EU. In 2015, EU multinationals registered a
tax deficit that ranged between a 1% to 4% of EU’s national income, depending on the minimum rate
specified. The corresponding share for non-EU MNEs ranged from 0.3% to 0.8%.
Which are the multinationals that register the highest tax deficit? From Figure 4, which
considers a minimum effective tax rate of 25%, among the EU, Germany, Netherlands and United
Kingdom MNEs are the ones with the highest deficit, representing approximately half of the total. Why
are German multinationals the ones with, by far, with the highest tax deficit? A potential explanation
is the low effective tax rate of Germany, of 11%38 (see Table 19 at the Appendix). As most of German
multinationals profits (approximately a 90%) are located at its territory of origin, given the difference
with the minimum effective tax rate considered of 25%, a high tax deficit is recorded. From Figure’s 4
bottom graph, excluding the EU, deficit is mostly due to foreign affiliates profits taxed below an effective
tax rate of 25% at the United States, Hong Kong, China or Singapore. The tax deficit allocated outside
the EU represents a 8% of the total, the rest located at EU countries39. Figure 5 top panel shows that
American, Swiss and Bermudian multinationals had the highest tax deficit within the EU representing a
60% of the tax deficit of non-EU multinationals. As per the bottom of Figure 5, it was mostly recorded
at Ireland, Netherlands and Luxembourg (approximately 55% of the total).
Tables 7 and 8 contain the estimated outcomes from a unilateral adoption of a 25% minimum
effective tax rate for France. As explained in Section 2, this implies that France adopts the global
minimum effective tax rate40, entirely collects the tax deficit of French controlled multinationals and38Tørsløv, Wier, and Zucman (2018) discuss the reasons why the effective tax rate in Germany is this low. One reason
is the distortion of corporate profits due to the inclusion of self-employed, that do not pay corporate taxes. Another
explanation is the gap between tax revenues as reported by the OECD (which is the source used to compute the 11% of
effective tax rate) vs. German national accounts. If national accounts data is considered, it raises to 14%. In addition, the
existence of different business taxes at German territory also contribute to the low effective tax recorded, as tax revenues
may fall under different categories (for example, general vs. local government).39From Table 6, the tax deficit of EU headquartered multinationals is approximately equal to €150 billion. Of this tax
deficit, as per authors’ computation, €12 billion, or 8%, was registered at non-EU countries.40In 2015, the corporate ETR for France was equal to 24%. See Table 19 at the Annex.
17
apportions the global deficit of non-French. Table 7 contains the tax deficit collected by France from
French multinationals for different minimum effective tax rates41. This deficit is mostly derived from
their profits in tax havens as Netherlands, Belgium, or Singapore and big economies as the United States
and the United Kingdom. In an scenario with an effective tax rate of 25%, the tax collected represents
approximately 0.3% of French national income for 2015. In the case case of 40%, the percentage raises
up to 0.8%.
Table 8 shows the extra-tax revenue that France could obtain by collecting the tax deficit
of non-French multinationals with economic activity within its territory, in an scenario with a 25% of
minimum ETR. Under a GFA scheme with sales, most of the tax revenue, approximately an 80%, is
derived from EU MNEs which sums up to €7 billion. In particular, from multinationals located in
Germany, Luxembourg, and Netherlands, as they have both a large share of sales located in France
and a global tax deficit. France could collect this deficit by setting an sales tariff of 1.3% on sales of
non-French EU multinationals. In other words, as a 5% of the total sales of EU multinationals are
registered in France, this country is entitled to recover a 5% of the global tax deficit or €7 billion.
Conversely, a tariff rate of 1.1% would be the one needed to raise the tax deficit of €3 billion from the
rest of the world. Note that this tariff is larger for tax haven multinationals as they record a higher
deficit. For example, the tariff in Bermudian multinationals for selling in France would be of 9.2%, or of
a 4.8% for Luxembourg. In 2017, the average tariff applied by France to all type of imported products
was 4.3%42. Therefore, the created tariff under this scheme would be lower than the existent ones.
Among the different GFA schemes, weighting by sales achieves the highest total tax deficit allocated to
France, of €9.3 billion. This is related with the fact (see Table 1) that sales are mostly concentrated at
high income countries, where final consumers are located. In total, under a GFA scheme with a 25% of
effective tax rate, France would collect approximately a 6% of the total tax deficit (see Table 7)43.
Considering a GFA scheme that includes the number of employees, total taxes collected range
between €8.7 to €9 billion (see Table 8). In the case of just defining employees as a factor, the tax
deficit could be collected by setting a per employee tax rate of 0.5% on non-French multinationals. In
other words, the multinational would have to pay €0.005 per each employee in French territory. Tables41If a simpler mechanism where to be adopted, where there is no distinction between French and non-French multina-
tionals tax deficit (see Section 2), in the case of considering sales, France will collect a 68% of the total tax deficit contained
at Table 7. Note that the general conclusion of this Section does not rely on which of the approaches is considered.42Source: World Trade Organization (WTO) Integrated Database (IDB). The average tariff rate of 4.3%, is computed
with the by product data in the case of trade partners that fall into the category of Most Favoured Nation (MFV) for
France.43Note that this value is computed for a total tax deficit that includes France, as it has adopted the minimum tax rate.
18
9 to 13 present the same information as Table 8 but considering 30%, 35%, 40%, 45% and 50% as
the minimum ETR. Due to the increase in the minimum rate, higher revenues are collected from EU
and non-EU multinationals but at the cost of an increase in sales tariff and per employee tax. For the
highest minimum ETR considered, 50%, the tax revenue collected by France increases to €21 billion
which could be achieved by applying a tariff rate of 4%.
From Figure 6, the associated tax revenues from a unilateral transition, in the case of France,
range between 0.7% to 3% of its national income, depending on the minimum rate defined. This result
is equivalent for all the GFA models considered. If the raise in taxation over domestic corporate profits
is also included, corporate tax revenues increase between €30 to €80 billion44 but, it does not create a
substantial change in national income terms.
But, does a country like France have the incentive to unilaterally adopt a minimum effective tax
rate? Or is it better-off if there is international cooperation to set a global minimum tax rate? To answer
this question, the expected tax revenue gains for France under each of these scenarios, cooperative and
unilateral, must be compared. To compute the tax revenue gains under a cooperative scenario, the
approach from Guvenen et al. (2017) is applied. Following their methodology, the resulting allocation
under a GFA scheme is defined as the location of profits in absence of shifting incentives. In a context
were there is international cooperation, the motivation to engage in profit shifting will disappear as there
are not tax differentials across countries (see Section 2). For the unilateral scenario, as aforementioned,
the resulting tax gains from collecting the tax deficit are shown at Table 7 and Tables 8 to 13 depending
on the minimum effective tax rate considered45. Note that, for comparability across both scenarios, we
just consider the tax revenues arising from EU multinationals’ profits.
Tables 14 to 16 contain the estimated tax revenues gains as compared with separate account-
ing for different GFA schemes definitions (sales, number of employees or both) and effective tax rates,
for both cooperative and unilateral scenario. In the case of defining sales as an apportionment factor,
independently from the effective tax rate specified, a country like France could have an incentive to
unilaterally implement a minimum effective tax rate (in the absence of retaliation strategies by other
countries). In the case of unilaterally fixing a tax rate of 35%, the tax revenue gains represent approx-
imately 2% of its national income, as compared with the 1.4% under a cooperative strategy. For the
maximum effective tax rate considered, a 50%, these percentages raise to 4% and 3%, respectively. The44These values are obtained from Table 19 at the Annex.45As mentioned at Section 2, the total tax revenue gains from this unilateral scenario also include the tax gains from
taxing the profits booked in France at the minimum ETR.
19
same phenomena arises when the GFA scheme is designed with the number of employees (Table 15) or
both sales and employees (Table 16). Therefore, we can conclude, that unilaterally adopting a minimum
effective tax rate under the system considered in this paper (see Section 2), is a tempting strategy for
a country like France due to its associated tax revenue gains.
6 Conclusions
This paper analyses the economic impact of the introduction of a Global Formulary Apportionment
(GFA) scheme at the European Union and the outcome associated with adopting a global minimum
effective tax rate. The main finding is that tax havens would be the main losers from a transition out of
the current international tax scheme. Profit shifting opportunities for multinationals would be reduced
with a system that changes taxing rights towards countries were economic activity is located and that
limits corporate tax strategic setting. This lost for tax havens is translated into sizeable benefits in
terms of tax base and revenues for all the different income groups, specially for countries at the mid
and the bottom of the income distribution or developing countries. For a country like France, adopting
unilaterally a minimum effective tax rate, could rise taxes collected between 0.7% and 4% of its national
income. In a scenario with cooperation where all countries set the same effective tax rate, tax revenue
gains increase by 0.6% to 3% of national income. Following these results, a country like France could
not have an incentive to cooperate towards fixing a minimum effective tax rate.
These findings have relevant policy implications for the re-design of international tax policy.
First, the evaluated reforms mostly benefit lower-middle- and low-income economies, which are more
reliant on corporate income tax revenues (Keen et al. (2014)). This could enable an increase in public
spending which may contribute to advances in their economic development (Forstater (2015)). This
resources may be crucial especially given the pessimistic forecasts of the economic aftermath following
the Coronavirus crisis. Second, the present levels of tax competition are predicted to raise in the near
future (IMF (2019)) which threatens the survival of the corporate income tax. Defining a minimum
global effective tax rate, could limit this competition. If international coordination is not available for
its implementation, countries could still adopt it and derive substantial tax revenues from it, as we have
observed for a country like France.
20
The presented results in this paper, face limitations especially concerning both the available
data and the absence of behavioural responses by firms and governments in the analysis. This paper
draws from publicly available data that faces challenges of harmonization across countries and coverage
for low income and tax haven countries. The OECD made an announcement that this last January
2020, aggregate statistics of multinationals would be available. As per when this paper was written,
these statistics are still not published. This data will contribute to obtain a more precise picture of the
policy implications of the reforms evaluated here.
21
7 References
Altshuler, Rosanne, and Harry Grubert. 2010. “Formula Apportionment: Is It Better Than the Current
System and Are There Better Alternatives?” National Tax Journal, December 63 (4): 1145–84.
Alvaredo, Facundo, Lucas Chancel, Thomas Piketty, Emmanuel Saez, and Gabriel Zucman. 2017. “The
Elephant Curve of Global Inequality and Growth.”
Avi-Yonah, Reuven S, and Kimberly A Clausing. 2007. “A Proposal to Adopt Formulary Apportionment
for Corporate Income Taxation: The Hamilton Project.” U of Michigan Law & Economics, Olin Working
Paper, nos. 07-009.
Beer, Sebastian, Ruud A Mooij, Shafik Hebous, Michael Keen, and Li Liu. 2020. “Exploring Residual
Profit Allocation.”
Chancel, Lucas. 2020. “Coronabonds with or Without Germany.” World Inequality Lab.
Clausing, Kimberly A. 2016. “The Us State Experience Under Formulary Apportionment: Are There
Lessons for International Reform?” National Tax Journal 69 (2): 353–86.
Cobham, Alex, Tommaso Faccio, and Valpy FitzGerald. 2019. “Global Inequalities in Taxing Rights:
An Early Evaluation of the Oecd Tax Reform Proposals.”
Cobham, Alex, and Petr Janskỳ. 2019. “Measuring Misalignment: The Location of Us Multinationals’
Economic Activity Versus the Location of Their Profits.” Development Policy Review 37 (1): 91–110.
Cobham, Alex, and Simon Loretz. 2014. “International Distribution of the Corporate Tax Base:
Implications of Different Apportionment Factors Under Unitary Taxation.”
De Mooij, Ruud A, Ms Li Liu, and Dinar Prihardini. 2019. An Assessment of Global Formula Appor-
tionment. International Monetary Fund.
Devereux, Michael P, and Simon Loretz. 2008. “The Effects of Eu Formula Apportionment on Corporate
Tax Revenues.” Fiscal Studies 29 (1): 1–33.
Eichfelder, Sebastian, Frank Hechtner, and Jochen Hundsdoerfer. 2018. “Formula Apportionment:
Factor Allocation and Tax Avoidance.” European Accounting Review 27 (4): 649–81.
Eurostat. 2012. Foreign Affiliates Statistics Recommendations Manual. European Commission Luxem-
22
bourg.
Forstater, Maya. 2015. “Can Stopping ‘Tax Dodging’ by Multinational Enterprises Close the Gap in
Development Finance?” CGD Policy Paper 69: 32.
Fuest, Clemens et al. 2007. “How Would the Introduction of an Eu-Wide Formula Apportionment Affect
the Distribution and Size of the Corporate Tax Base? An Analysis Based on German Multinationals.”
International Tax and Public Finance 14 (5): 605–26.
Fuest, Clemens, Mathieu Parenti, and Farid Toubal. 2019. “International Corporate Taxation: What
Reforms? What Impact?” Notes Du Conseil d’analyse économique, no. 6: 1–12.
Garcia-Bernardo, Javier, Petr Jansky, and others. 2019. “Multinational Corporations and Tax Havens:
Evidence from Country-by-Country Reporting.” Charles University Prague, Faculty of Social Sciences,
Institute of Economic ….
Glassman, Amanda. 2020. “Coronavirus and Low-Income Countries: Ready to Respond?”
Guvenen, Fatih, Raymond J Mataloni Jr, Dylan G Rassier, and Kim J Ruhl. 2017. “Offshore Profit
Shifting and Domestic Productivity Measurement.” National Bureau of Economic Research.
Hines Jr, James R. 2010. “Income Misattribution Under Formula Apportionment.” European Economic
Review 54 (1): 108–20.
ICRICT. 2018. “A Roadmap to Improve Rules for Taxing Multinationals. A Fairer Future for Global
Taxation.”
IMF. 2019. Corporate Taxation in the Global Economy. IMF.
Keen, MJ, Victoria Perry, R d Mooij, Thornton Matheson, Roberto Schatan, Peter Mullins, and Ernesto
Crivelli. 2014. Spillovers in International Corporate Taxation. IMF.
Laffitte, Sébastien, and Farid et al. Toubal. 2019. “A Fistful of Dollars? Foreign Sales Platforms and
Profit Shifting in Tax Havens.”
OECD. 2005. OECD Handbook on Economic Globalisation Indicators. OECD.
———. 2009. OECD Benchmark Definition of Foreign Direct Investment 2008. Organisation for
Economic Co-operation; Development.
23
———. 2016a. “Foreign Direct Investment Statistics - Explanatory Notes.” Organisation for Economic
Co-operation; Development - Investment Division.
———. 2016b. “Glossary of Foreign Direct Investment Terms and Definitions.” Available at:
* Authors’ calculations from FDI data (Eurostat and OECD), National Accounts (NASA) and Foreign AffiliatesStatistics data (Eurostat and OECD). * Values in parenthesis correspond to the share with respect to the total.
30
Table 2: Profit allocation under Global Formulary Apportionment(GFA) by income level, bn EUR (2015)
1 Source: Authors’ calculations from FDI data (Eurostat and OECD), National Accounts (NASA) and Foreign Affiliates Statistics data (Eurostat and OECD).2 Initial refers to the current allocation of the tax base under a separate accounting system. The difference between total initial allocation and for the different GFA
(€25 billion), is due to multinational profits being consolidated under the GFA system.3 The total value for the relative gain (% gain), corresponds to the average relative gain across income groups.4 Income classification is from World Bank.
31
Table 3: Profit allocation under Global Formulary Apportionment(GFA) by development, bn EUR (2015)
1 Source: Authors’ calculations from FDI data (Eurostat and OECD), National Accounts (NASA) and Foreign Affiliates Statistics data (Eurostat and OECD).2 Initial refers to the current allocation of the tax base under a separate accounting system. The difference between total initial allocation and for the differnt GFA
(€25 billion), is due to multinational profits being consolidated under the GFA system.3 The total value for the relative gain (% gain), corresponds to the average relative gain across development groups.4 Development classification is from UNCTAD.
32
Table 4: Estimated corporate tax revenue per income level, bn EUR(2015)
Sep. Accounting GFA Sales GFA Employment GFA Sales and empl. GFA Population% NI Per
1 Source: Tax revenues are estimated with statutory corporate income tax from KPMG and the World Bank. Data for national income and population is from the WorldInequality Database (WID).
33
Table 5: Estimated corporate tax revenue per development level, bnEUR (2015)
Sep. Accounting GFA Sales GFA Employment GFA Sales and empl. GFA Population% NI Per
1 Source: Tax revenues are estimated with statutory corporate income tax from KPMG and the World Bank. Data for national income and population is from theWorld Inequality Database (WID).
34
Table 6: Estimated tax deficit by multinational’s territory of origin, mn EUR (2015)
European Union 148.9 1.2 223.5 1.8 298.4 2.4 373.5 3.0 448.7 3.6 524.0 4.3Rest of the world 33.0 0.3 44.9 0.4 56.8 0.5 68.7 0.6 80.7 0.7 92.8 0.8Total 181.9 3.0 268.4 4.4 355.2 5.8 442.3 7.2 529.3 8.6 616.8 10.01 Source: Authors’ calculations from FDI data (Eurostat and OECD), National Accounts (NASA), World Inequality Database (WID) and effective tax
rates from Torslov et al. (2018) or imputed.2 National Income (NI) is the total for the European Union for 2015.3 The tax deficit corresponds to the tax revenue not collected at a certain jurisdiction due to profits not being taxed at a minimum pre-specified ETR (in
this case: 25%, 30%, 35%, 40%, 45% and 50%).
35
Table 7: Tax deficit of French controlled multinationals per countryand minimum ETR, mn EUR (2015)
rates from Torslov et al. (2018) or imputed.2 The tax deficit corresponds to the tax revenue not collected at a certain jurisdiction due to profits not being taxed
at a minimum pre-specified ETR.
36
Table 8: Tax deficit collected by France from non-French multination-als with a minimum ETR of 25, mn EUR (2015)
Table 14: Tax revenue gains for France from adopting a minimumETR apportioning by sales, mn EUR (2015)
Minimum ETR (%) 25% 30% 35% 40% 45% 50%
Cooperative scenario 11184.8 19311.9 26727.8 34144.32 41560.8 48977.4% national income 0.6 1.0 1.4 1.80 2.2 2.6Unilateral scenario 13447.8 24935.8 36479.5 42371.90 59568.3 71117.2% national income 0.7 1.3 1.9 2.20 3.1 3.81 Source: Authors’ calculations from FDI data (Eurostat and OECD), National Accounts (NASA)2 Under separate accounting, the profits allocated to France from EU multinationals are €105 billion. The total gain for the
unilateral scenario includes the tax deficit collected and the tax revenue gain from taxing these profits at the minimum ETR.3 Gains for the cooperative scenario are computed with respected to the the tax revenues under a separate accounting tax system.
The tax revenues are estimated with France ETR for 2015, 24%4 For comparability, the tax revenues under the unilateral scenario just includes the corporate profits of EU multinationals.
Table 15: Tax revenue gains for France from adopting a minimumETR apportioning by the number of employees, mn EUR (2015)
Minimum ETR (%) 25% 30% 35% 40% 45% 50%
Cooperative Scenario 2806.1 8404.9 14003.6 19602.4 25201.2 30800.0% national income 0.1 0.4 0.7 1.0 1.3 1.6Unilateral scenario 12979.3 24274.5 35617.3 46974.1 58331.0 69687.9% national income 0.7 1.3 1.9 2.5 3.1 3.71 Source: Authors’ calculations from FDI data (Eurostat and OECD), National Accounts (NASA)2 See Table 12 footnotes.
Table 16: Tax revenue gains for France from adopting a minimumETR apportioning by sales and the number of employees, mn EUR(2015)
Minimum ETR (%) 25% 30% 35% 40% 45% 50%
Cooperative scenario 7351.6 13859.5 20367.4 26875.3 33383.2 39891.1% national income 0.4 0.7 1.1 1.4 1.8 2.1Unilateral scenario 18831.6 24605.1 36043.9 47497.0 58949.6 70402.5% national income 1.0 1.3 1.9 2.5 3.1 3.71 Source: Authors’ calculations from FDI data (Eurostat and OECD), National Accounts (NASA)2 See Table 12 footnotes.
43
8 Annex8.1 Additional Tables
Table 17: Net national income and corporate profits by income leveland population, bn EUR (2015)
Total 54855.7 1495.9 7.2 72239.9 6514.41 Net national income (at market prices) and population statistics from the World Inequality Database (WID). Total
value for per capita income and corporate profits corresponds to the average for all income groups.2 Note that the data for population just includes the set of countries where European Multinationals have registered
corporate profits.
44
Table 18: Country classification by income level and tax haven
High income Australia - Austria - Bouvet Island (NO) - British Indian Ocean Territory (UK) - Brunei Darussalam - Canada - Chile - ChristmasIsland (AU) - Cocos (Keeling) Islands (AU) - Cook Islands (NZ) - Croatia - Czechia - Denmark - Estonia - Falkland Islands (UK) -Faroes (DK) - Finland - France - French Polynesia (FR) - French Southern Territories (FR) - Germany - Greece - Greenland (DK) -Guam (US) - Heard Island and McDonald Islands (AU) - Hungary - Iceland - Israel - Italy - Japan - Kuwait - Latvia - Lithuania -Montserrat (UK) - Nauru - New Caledonia (FR) - New Zealand - Niue (NZ) - Norfolk Island (AU) - Northern Mariana Islands (US) -Norway - Oman - Pitcairn Islands (UK) - Poland - Portugal - Qatar - Saint Helena, Ascension and Tristan da Cunha (UK) - San Marino- Slovakia - Slovenia - South Georgia and the South Sandwich Islands (UK) - South Korea - Spain - Sweden - Taiwan - Tokelau (NZ) -Trinidad and Tobago - United Arab Emirates - United Kingdom - United States - United States Minor Outlying Islands - Uruguay -Vatican City State - Wallis and Futuna (FR)
Upper middle income Albania - Algeria - American Samoa (US) - Angola - Argentina - Armenia - Azerbaijan - Belarus - Bosnia and Herzegovina - Botswana -Brazil - Bulgaria - China - Colombia - Costa Rica - Cuba - Dominica - Dominican Republic - Ecuador - Equatorial Guinea - Fiji - Gabon- Georgia - Guyana - Iran - Iraq - Jamaica - Jordan - Kazakhstan - Libya - Malaysia - Maldives - Mexico - Montenegro - Namibia -North Macedonia - Palau - Paraguay - Peru - Romania - Russia - Serbia - South Africa - Suriname - Thailand - Turkey - Turkmenistan -Tuvalu - Venezuela
Lower middle income Bangladesh - Bhutan - Bolivia - Cabo Verde - Cambodia - Cameroon - Congo - Côte d’Ivoire - Djibouti - Egypt - El Salvador - Eswatini- Federated States of Micronesia - Ghana - Guatemala - Honduras - India - Indonesia - Kenya - Kiribati - Kosovo - Kyrgyzstan - Laos -Lesotho - Mauritania - Moldova - Mongolia - Morocco - Myanmar/Burma - Nicaragua - Nigeria - Pakistan - Palestine - Papua NewGuinea - Philippines - São Tomé and Príncipe - Saudi Arabia - Solomon Islands - Sri Lanka - Sudan - Syria - Tajikistan - Timor-Leste -Tonga - Tunisia - Ukraine - Uzbekistan - Vietnam - Yemen - Zambia
Low income Afghanistan - Benin - Burkina Faso - Burundi - Central African Republic - Chad - Comoros - Democratic Republic of the Congo -Eritrea - Ethiopia - Gambia, The - Guinea - Guinea-Bissau - Haiti - Liberia - Madagascar - Malawi - Mali - Mozambique - Nepal - Niger- North Korea - Rwanda - Senegal - Sierra Leone - Somalia - South Sudan - Tanzania - Togo - Uganda - Zimbabwe
Tax haven Andorra - Anguilla (UK) - Antigua and Barbuda - Aruba (NL) - Bahamas - Bahrain - Barbados - Belgium - Belize - Bermuda (UK) -Bonaire, Saint Eustatius and Saba - British Virgin Islands (UK) - Cayman Islands (UK) - Curaçao - Cyprus - Former NetherlandsAntilles - Gibraltar (UK) - Grenada - Guernsey - Hong Kong - Ireland - Isle Of Man - Jersey - Lebanon - Liechtenstein - Luxembourg -Macao - Malta - Marshall Islands - Mauritius - Monaco - Netherlands - Panama - Puerto Rico - Saint Kitts and Nevis - Saint Lucia -Saint Vincent and the Grenadines - Samoa - Seychelles - Singapore - Sint Maarten - Switzerland - Turks and Caicos Islands (UK) - USVirgin Islands (US) - Vanuatu
1 Income level classification from the World Bank - World Development Indicators (2015). Tax haven classification is from Torslov et al. (2018).
45
Table 19: Domestic corporate profits for EU countries, bn EUR (2015)
Total 3467.6 -197.4 1512.3 1095.0 689.7 367.9 2152.7 121.3 2274.0 332.3 14.61 Source: Non-financial Annual Sector Accounts (NASA) from Eurostat. Corporate tax revenues data is from OECD and IMf WoRLD. Data for missing profits is from
Torslov et al. (2018).2 Data for Croatia for 2015 is inferred with available data from previous years.
46
Table 20: Domestic corporate profits for EU countries divided by for-eign and local corporations, bn EUR (2015)
Total 2274.0 2152.8 121.2 481.1 285.6 132.3 63.2 1792.9 895.81 Non-financial Annual Sector Accounts (NASA) from Eurostat. Data for for domestic foreign corporate profits is from OECD and Eurostat inward equity income FDI.
Data for missing profits is from Torslov et al. (2018).2 Data for Croatia for 2015 is inferred with available data from previous years.3 Data for local multinationals is imputed Internal Revenue Service (IRS) and Torslov et al. (2018) data.
47
Table 21: Reported equity income on outward FDI vs. inward FDIreported by OECD and EU partners, bn EUR (2015)
Total 301.5 321.3 19.6 3.3 2.8 -0.51 Source: Eurostat and OECD FDI statistics.2 Spain and Luxembourg do not report data on equity income at a by partner country level.3 Data for Austria is for Resident Operating Units (non-SPE). Austria treats SPEs as confidential and as a
consequence data for all resident units its non-publishable.
48
Table 22: Bilateral discrepancies in outward FDI equity income bypartner country, mn EUR (2015)
Total 575.4 414.3 291.6 161 288.6 124.61 Source: Eurostat and OECD FDI statistics.2 Spain and Luxembourg do not report data on equity income at a by partner country level.3 Data for Austria is for Resident Operating Units (non-SPE). Austria treats SPEs as confidential and as a consequence
data for all resident units its non-publishable.
50
Table 24: Rate of return on FDI for all resident units, in percentage(2015)
Inward OutwardInward FDI Outward FDI Gap (outward - inward)
Total Equity Debt Total Equity Debt Total Equity Debt
Total 414.2 545.7 131.4 2.9 6.3 3.4 539.51 Source: Eurostat and OECD FDI statistics.2 Spain and Luxembourg do not report data on equity income at a by partner country level.3 Data for Austria is for Resident Operating Units (non-SPE). Austria treats SPEs as confidential and as a consequence data
for all resident units its non-publishable.
52
Table 26: Impact of unallocated or confidential sales and employeeson outward FATS data (2015)
Sales (bn. EUR) Employees (th.)World By partner Unallocated
Total 6886.6 6838.2 350.5 24157.7 23296.0 1520.61 Source: Foreign Affiliate Statistics (FATS) from Eurostat and Activity of Multinational Enterprises (AMNE) from
OECD.2 For Bulgaria, Estonia and Netherlands the total value for world is not reported.
53
Table 27: Reported sales and number of employees for outward vs.inward FATS reported by EU and OECD partners (2015)
Total 6062.4 4801.6 -1260.7 12709.6 12860.4 150.81 Source: Foreign Affiliate Statistics (FATS) from Eurostat and Activity of Multinational Enterprises (AMNE)
from OECD.
54
Table 28: Corrected sales and employees for outward FATS (2015)
Sales (bn EUR) Employees (th.)Reported Corrected Gap
Total 6838.2 8855.6 2017.4 23296.0 27927.3 4631.31 Source: Foreign Affiliate Statistics (FATS) from Eurostat and Activity of Multinational Enterprises (AMNE)
from OECD.
55
8.2 Difference between headquartered and resident corporations at the EuropeanUnion
An interesting insight from Foreign Affiliate Statistics (FATS) (see Section 3 for more details) is that
statistics are presented by ultimate controlling parent. More concretely, by the country of location of
the corporation headquarters. This allows to infer where corporations of a certain nationality locate
their economic activity. Tables 29 and 30 compare the shares of sales and employees, of corporations
headquartered in a country with respect to the total of resident corporation at its territory for European
Union corporations. For example, from Table 29, French corporations registered a total of €4.3 trillion
worldwide of which 65% where made in France. Resident corporations in France, independently of
whether they are headquartered or not in its territory, gained around €4 trillion in sales. Therefore,
French corporations generated more income from its sales than the total of corporations at its territory by
a factor of 1.2. If we compare across all countries, a stylized fact arises, Eastern European corporations
generate less income from its activity outside its frontier than the rest of EU corporations. This is
indicative of a lower international presence of Eastern European corporations as it is observed in its
levels of equity income on outward FDI (see Table 25). As logic, the same pattern is observed in the case
of employees at Table 30. Corporations from richer EU countries register a higher share of employees
outside its frontier with respect to the total of resident corporations.
56
Table 29: Sales of resident vs. controlled European Union corpora-tions, bn EUR (2015)
Controlled by Resident atTotal Foreign % Domestic % Total Controlled
Total 26561.8 7188.7 27.06 19373.1 72.94 27269.9 97.40
Note:Foreign corresponds to the total of sales of a corporation outside the frontier where it is headquartered. In thecase of domestic, they are the sales by corporation within the territory where they are headquartered.
1 Source: Foreign Affiliate Statistics (FATS) and Structural Business Statistics (SBS) from Eurostat.
57
Table 30: Persons employed for resident vs. controlled of EuropeanUnion corporations, in thousands (2015)
Controlled by Resident atTotal Foreign % Domestic % Total Controlled
Total 139401.2 24816.5 17.80 114584.6 82.20 137371.5 101.48
Note:Foreign corresponds to the total number of employees of a corporation outside the frontier where it is headquar-tered. In the case of domestic, they are the number of employees of the corporation within the territory wherethey are headquartered.
1 Source: Foreign Affiliate Statistics (FATS) and Structural Business Statistics (SBS) from Eurostat.
